SPACECAP(SPACECAP)
SPACECAP()所属R语言包：SPACECAP
A Program to Estimate Animal Abundance and Density using Bayesian Spatially-Explicit Capture-Recapture Models
一个程序，使用贝叶斯空间显式捕获 - 再捕获模型的估计动物的丰度和密度
译者：生物统计家园网 机器人LoveR
描述----------Description----------
SPACECAP is a user-friendly software package for estimating animal densities using closed model capture-recapture sampling specifically based on photographic captures. It implements a new generation of spatially explicit capture-recapture models developed in a recent paper, Royle et al.(2009), cited below. Spatially explicit capture-recapture models implemented in SPACECAP directly estimate animal density by explicitly using the information on capture histories in combination with spatial locations of captures under a unified Bayesian modeling framework. This approach offers advantage such as: substantially dealing with problems posed by individual heterogeneity in capture probabilities in conventional capture-recapture analyses. It also offers non-asymptotic inferences which are more appropriate for small samples of capture data typical of photo-capture studies.
SPACECAP是一个用户友好的软件使用封闭模型捕获 - 再捕获抽样专门根据摄影捕捉动物密度估计的包。它实现了新一代的空间明确的捕获 - 再捕获模型开发在最近的一篇文章，罗伊尔等。人（2009年），列举如下。空间明确的捕获 - 再捕获模型SPACECAP直接估算动物的密度，结合空间位置捕获一个统一的贝叶斯模型框架下通过显式使用的信息采集历史。这种方法提供了优势，如：基本上与个体异质性所带来的问题，在捕获概率在传统的捕获 - 再捕获分析处理。此外，还提供了非渐进的推论，更适合小样本的采集数据，典型的照片捕捉的研究。
用法----------Usage----------
SPACECAP()
Details
详细信息----------Details----------
PROGRAM SPACECAP Version 1.0.6
PROGRAM SPACECAP版本1.0.6
Introduction
介绍
SPACECAP, is a user-friendly software package for estimating animal densities using closed model capture-recapture sampling specifically based on photographic captures. It implements a new generation of spatially explicit capture-recapture models developed in a recent paper titled "Bayesian Inference in Camera Trapping Studies for a Class of Spatially Explicit Capture Models" by J. A Royle, K. U. Karanth, A. M. Gopalaswamy, and N. S. Kumar, published in Ecology, 90(11), 2009, pp. 3233-3244.
SPACECAP，估计使用封闭模型捕获 - 再捕获抽样专门根据摄影捕捉的动物密度是一个用户友好的软件包。它实现了新一代的空间明确的捕获 - 再捕获模型开发在最近的一篇文章，题为“贝叶斯推理的空间显式捕获模型的一类相机中捕获研究”由J.罗伊尔，KU Karanth，AM Gopalaswamy，NS库马尔，发表在生态学报，90（11），2009年，第3233-3244。
The conventional approach to the analysis of animal density from camera trap surveys is to apply closed capture-recapture model analyses, and, then convert resulting estimates of abundances to densities using a wide range of essentially ad hoc methods. For example, a standard practice in camera trapping studies has been to use auxiliary capture location information to estimate the mean or maximum distance moved by the study species as a basis to apply a "buffer width" around the trap array or, to use various other heuristic "adjustments" (Wilson and Anderson 1985a, b, Karanth and Nichols 1998, Parmenter et al. 2003, Trolle and Kery 2003) to estimate the effectively sampled area. While these approaches appear to work adequately in practice, little had been known about the range of conditions under which they work well. This is because most real world study situations involve study area of odd shapes and sizes and difficult terrain that makes setting camera traps challenging and conditions assumed by ad hoc approaches may not apply.
动物的密度从相机陷阱调查分析，传统的方法是申请封闭捕获 - 再捕获模型分析，然后将其转换导致的丰度估计密度使用范围广，基本上是专案方法。例如，一个标准的做法在相机捕获的研究一直使用辅助的拍摄位置信息来估计的平均或最大的移动距离由研究物种作为基础申请一个“缓冲宽度”周围的陷阱阵列，或使用各种其他启发式“调整”（威尔逊和安德森1985A，B，Karanth和Nichols在1998年，帕门等。2003年，Trolle和柯瑞2003）估算的有效样本区域。虽然这些方法似乎是在实践中充分，小的条件下，他们的工作的范围。这是因为，全球最真实的学习情况涉及研究领域的奇怪的形状和大小，复杂的地形，使相机陷阱的挑战和承担专案的方法可能并不适用条件。
Spatially explicit capture-recapture models implemented in SPACECAP package directly estimate animal density by explicitly using the information on capture histories in combination with spatial locations of captures under a unified Bayesian modeling framework. This approach offers advantage such as: substantially dealing with problems posed by individual heterogeneity in capture probabilities in conventional capture-recapture analyses. It also offers non-asymptotic inferences which are more appropriate for small samples of capture data typical of photo-capture studies. Further technical details about the models and analyses can be obtained from the paper by Royle et al. (2009), cited above. We also note another spatially explicit modeling and estimation approach based on conventional likelihood based inference that is available (Borchers and Efford 2008).
通过显式使用的信息捕捉历史上，结合空间位置捕获一个统一的贝叶斯模型框架下，空间明确的实施SPACECAP包捕获 - 再捕获模型直接估算的动物的密度。这种方法提供了优势，如：基本上与个体异质性所带来的问题，在捕获概率在传统的捕获 - 再捕获分析处理。此外，还提供了非渐进的推论，更适合小样本的采集数据，典型的照片捕捉的研究。进一步的技术细节的模型和分析，可以从纸张的罗伊尔等。 （2009年），上面提到。我们还注意到另外一个直观的建模和估计方法的基础上常规的可能性的推理是（BORCHERS和Efford 2008）。
The Spatial Capture-Recapture Model and its Parameters
捕获 - 再捕获的空间模型及其参数
The model considered in the current version of SPACECAP applies to binary observations y(i,j,k) for individual "i", trap "j" and sample occasion (e.g., night) "k". The model is a type of binary regression model, similar to logistic regression, in which
该模型认为在当前版本的SPACECAP为二进制观测Y（I，J，K）为单独的“我”，陷阱“J”和样品的场合（如晚上），“K”。该模型是一个二进制类型的回归模型，类似logistic回归，在其中
Here p(i, j, k) is the probability of detecting an individual "i", at trap "j" and sample occasion "k". The probability p(i,j,k) is then related to suitable covariates of interest by applying a suitable transformation. In SPACECAP we make use of the complementary log-log link transformation. Thus, the simplest possible model is expressed
这里p（I，J，k）的是个人的“i”，在阱“j”和样品的场合“k”的检测的概率。的概率p（I，J，k）的，然后通过施加合适的变换有关合适的协变量的兴趣。在SPACECAP的“我们利用的互补loglog链接转型。因此，可能的最简单的模型来表示
where b0 is a parameter to be estimated. By taking the inverse of the cloglog transformation
其中b0是要估计的参数。通过采取的cloglog变换的逆
This particular function arises by considering the binary observations to be formally reduced from a Poisson encounter frequency model (see Royle et al 2009 for details). That is, p(i,j,k) is the Pr(at least 1 encounter in a trap) under the Poisson model. The most general model that SPACECAP presently allows is a model in which
这种特殊的功能产生，正式从泊松遭遇频率模式（见2009年罗伊尔等）的二进制意见。也就是说，P（I，J，K）是泊松模型下的Pr（至少1中遇到的一个陷阱）。最普遍的模式，SPACECAP目前允许一个模型，其中
In this expression x(i,j,k) is an indicator of previous capture of individual i in trap j. Thus "b1" is a measure of the behavioral response (see below for more discussion of this; also see Royle et al 2009 for more details). The parameter b2 is a regression coefficient on the effect of distance between individual activity center s(i) and the location of trap j, u(j). b2 is constrained to be < 0 implying that the probability of encountering an individual in a trap decreases as the distance between the individual and the trap increases. In the model, s(i) are the collection of animal locations within a prescribed region S (this is provided by the user - see below). S is referred to subsequently as the state-space of the random variables s(i).
在这个表达式X（I，J，K）是以前的捕获的陷阱Ĵ我个人的指标。因此，“B1”是衡量行为反应（见下文详细讨论;有关详细信息，请参阅2009年罗伊尔等）。参数b2是回归系数对个别活动中心（i）和陷阱J，U（j）条的位置之间的距离的效果。 b2的约束为<0这意味着在一个陷阱遇到个人的概率作为个人和陷阱之间的距离增加而减小。在此模型中，（i）是在规定的区域S（这是由用户提供的 - 见下文）的动物的位置的集合。 S被称为随后作为随机变量s（ⅰ）的状态空间。
In SPACECAP, the parameter N is the population size of individuals - the number of activity centers located in S (see further discussion of this below).
在SPACECAP，参数N是个人的人口规模 - 在S活动中心（见下面进一步讨论）。
In SPACECAP, N is reported under the name Nsuper. <br> Density, D = N/||S|| where ||S|| is the area of the state-space. The units for Density, D is animals/100 km2.
在SPACECAP，N据悉的名义下Nsuper。 <BR>密度，D = N / | | S | |哪里| | S | |是状态空间的面积。单位的密度，D是animals/100平方公里。
SPACECAP also reports several derived parameters:
SPACECAP还报告了几个导出参数：
lam0 = exp(b0), is the intercept in terms of expected encounter frequency; this may be thought of as the expected encounter rate of an individual "i" in trap location "j" at sampling occasion "k", whose home-range centre is exactly at the trap location.
lam0 =进出口（b0）中的是截距在预期遭遇次数方面，这可以被认为是作为一个单独的“i”的陷阱位置“j”的“k”的在采样的场合，其预期遭遇率家庭范围的中心正是在陷阱位置。
sigma = sqrt(1/b2), is the scale parameter of a bivariate normal encounter function (or an exponential). This may also be viewed as a "range parameter" of an animal. For a highly mobile animal, this value will tend to be large (eg:- tigers will have a higher sigma value in comparison to civets)
σ=的SQRT（1/b2），是规模参数的二元的正常相遇功能（或指数）。这也可以看作是一个“范围参数”的动物。对于高移动性的动物，这个值会趋向于大（例如： - 老虎将有一个更高的西格玛值相比，果子狸）
beta = b1, is the regression coefficient that measures the behavioral response
β= B1，是回归系数测量的行为反应
psi = the ratio of the number of animals actually present within S to the maximum allowable number (set by the user during data augmentation - see below).
psi的=实际存在于S到的最大允许数量（数据扩张期间由用户设定 - 见下文）的动物的数量的比率。
Installing PROGRAM SPACECAP Version 1.0.6
安装程序SPACECAP版本1.0.6
STEP 1: Download latest version of program R (R Development Core Team)
第1步：下载最新版本的程序R（R发展的核心团队）
SPACECAP works within the R programming environment, so your very first step will be to connect to the internet, go to the website http://www.r-project.org, download and install the latest version of R (R 2.15.0 or higher) from the nearest CRAN mirror.
SPACECAP内的R编程环境，所以你的第一个步骤将是连接到互联网上，去网站，http://www.r-project.org，下载并安装最新版本的R（ ŕ2.15.0或更高版本）从最近的CRAN镜。
STEP 2: Download the package SPACECAP to your computer.
第2步：下载包SPACECAP到您的计算机。
When you launch R, go to Packages->Install package(s), once again select your nearest CRAN mirror and select package SPACECAP for installation.
当您启动R，去包 - >安装包（s），再次选择离你最近的CRAN镜，并选择套件“SPACECAP安装。
STEP 3: OPEN the package SPACECAP
步骤3：打开包装SPACECAP
In the R environment, go to Packages->Load package, select SPACECAP and load the package.
在R环境中，去包 - >加载包，选择SPACECAP和加载包。
STEP 4: Launching SPACECAP
第4步：启动SPACECAP
In the R environment, at the prompt ">", type the command SPACECAP(). This will launch the Graphic-User Interface of SPACECAP and you are now set to begin the Bayesian Spatially-Explicit Capture-Recapture (SECR) analysis of your camera trap survey data.
在R环境中，在提示符“>”，键入命令SPACECAP的（）。这将启动图形用户界面的SPACECAP和贝叶斯空间显式捕获 - 再捕获（SECR）你的相机陷阱调查数据的分析，你现在开始。
SECR Analysis using SPACECAP
SECR分析中的应用SPACECAP
Running an SECR Analysis in SPACECAP essentially involves four simple steps:
运行的SECR分析SPACECAP主要涉及四个简单的步骤：
1. Setting up the input files <br> 2. Selecting the appropriate model combination <br> 3. Selecting the Markov chain Monte Carlo (MCMC) settings <br> 4. Hitting the "RUN" button. <br>
1。设置输入文件参考。选择合适的模型组合参考3。选择马尔可夫链蒙特卡罗（MCMC）设置参考4。点击“RUN”按钮。参考
STEP 1: SETTING UP THE INPUT FILES FOR ANALYSIS
第1步：设置输入文件进行分析
SPACECAP requires you to first create three input files and store these files at a suitable location on your computer. These files are:
SPACECAP要求您先创建三个输入文件，将这些文件存储在您的计算机在一个合适的位置。这些文件是：
1. Animal Capture Details File (Animal ID no., Trap Location no., Sampling Occasion no.)
1。动物捕捉细节文件（动物ID。，陷阱位置，采样场合。）
2. Trap Deployment details File (Trap Spatial Location, Deployment Activity, Sampling Occasion no.)
2。陷阱部署详细资料档案（陷阱的空间位置，部署活动，采样场合。）
3. State-space details File (describing the Potential Animal Home Range Center Details in terms of their spatial location and habitat suitability indicator for these home range centers)
3。国家空间的详细信息文件（描述了潜在的动物的活动范围在它们的空间位置和栖息地适宜度分析指标，为这些家庭范围中心的中心详情）
These three raw data files can most easily be created using spreadsheet applications like Microsoft EXCEL, OpenOffice or other software you are comfortable with. However, eventually, all files must be saved in an ASCII comma separated format (.csv), because SPACECAP can only read these types of input files.
这三个原始数据文件，可以很容易地使用电子表格应用程序，如Microsoft Excel，OpenOffice或其他软件，你是舒适的。然而，最终，所有的文件必须保存在ASCII逗号分隔的格式（CSV），这是因为SPACECAP只能读这些类型的输入文件。
Upon launching SPACECAP, you will notice on the input data panel, three separate buttons to load the three input data files. Pressing of these buttons will help you locate the corresponding input files using your desktop browser (like Windows Explorer or Finder).
在启动SPACECAP，你会发现在输入数据“面板中，三个独立的按钮，装载三个输入数据文件。按这些按钮将帮助您找到相应的输入文件使用的桌面浏览器（如Windows资源管理器或Finder）。
INPUT FILE 1: Animal Capture Details
INPUT FILE 1：动物捕捉细节
The input file containing individual animal capture histories and locations consists of a 3-column table, each column representing the Location Number, the Animal Identity Number and the Sampling Occasion number, strictly in that order. Please note that these are all "number" fields. Please not enter labels containing alpha-numeric characters such as "CAT-123", "PLACE-100" or "January 2009" etc. for these fields, because SPACECAP will not recognize them. Use simple integer numbers. Each unique individual captured during camera trap sampling should be given a unique identification number, ranging from 1 to n, where n is the total number of unique individuals caught during the camera trap survey.
输入文件，其中包含个人的动物捕获的历史和位置，包括一个3列的表格，每一列代表的位置编号，动物身份识别号码和采样场合，严格的顺序。请注意，这些都是“数字”字段。请输入标签包含字母数字字符，如“CAT-123”，“PLACE-100”或“2009年1月”等这些领域，因为SPACECAP将无法识别它们。使用简单的整数。捕获相机陷阱抽样过程中每一个独特的个体，应给予一个唯一的标识号，取值范围从1到n，其中n是夹在相机陷阱调查总数的独特的个人。
Duration of the overall survey is determined by species biology in order to meet the assumption of demographic closure. The duration of sampling occasions (or periods) will, in turn, be based on how many such occasions are there in the survey duration.
的综合调查是由物种生物学的时间，以满足假设的人口关闭。反过来，根据持续时间的的采样场合（或期间），在调查期间，有多少这样的场合。
Because of shortage or camera traps or logistical reasons in many camera trap surveys, the study area is quite often divided into blocks or sub-units and camera trapping is conducted sequentially in these (Karanth and Nichols 2002). With data from conventional capture-recapture analyses the resulting capture history matrix is artificially constructed to ensure that assumption that camera trap survey simultaneously covers the entire area on each sampling occasion (See Karanth and Nichols 2002 page 132-133).
由于短缺或的相机陷阱或后勤方面的原因，在许多相机陷阱调查，研究区域划分成块或子单位，并在这些摄像头捕获顺序进行经常（2002年Karanth和Nichols）。与从传统的捕获 - 再捕获分析的数据的所捕获的历史矩阵人工构建的，以确保该假设，相机陷阱调查同时覆盖的整个区域，在每个采样的场合（见Karanth和Nichols 2002页132-133）。
However, since spatial locations of traps are explicitly incorporated into capture-recapture modeling in SPACECAP this artificial construction of sampling occasions is not required, which greatly enhances flexibility in survey and analyses. Each sampling occasion must have a unique identity number, ranging from 1: T, where T is the total number of sampling occasions. We note that each sampling occasion need not cover the entire survey area.
然而，由于空间位置的陷阱捕获 - 再捕获模型中明确纳入SPACECAP这种人工建造的样本的不是必需的，这大大增强了灵活性的调查和分析。每个采样的场合，必须有一个唯一的标识号，从1：T，其中T为样本的总数。我们注意到每一个的采样场合需要覆盖整个调查区域。
Each camera trap location must be given a unique identification number, ranging from 1: J, where J is the total number of camera trap locations used in the survey.
每个相机陷阱位置必须给予一个唯一的标识号，从1：J，其中J是总人数的调查中使用的相机陷阱的位置。
Assuming a camera trapping survey was conducted from 10 Jan 2009 to 30 Jan 2009, and, we treat each day as a sampling occasion, we end up with 20 sampling occasions. Let us assume there were 16 camera trap locations used in this survey and for logistical reasons the study area was partitioned into 4 blocks and each block contained 4 camera trap locations. Let us assume further that camera traps were deployed for 5 successive days in each block covering each of the four blocks successively in 20 days. Assuming only 6 animals were photo-captured and identified, the INPUT FILE 1 for SPACECAP would look like:
假设一个摄像头捕获的调查，2009年1月10日至2009年1月30日，我们把每一天都当作一个采样场合，我们结束了20个采样场合。让我们假设有16个摄像机陷阱在本次调查中使用的位置和后勤方面的原因，研究区域划分为4块，每块包含4个相机陷阱的位置。让我们进一步假设，相机陷阱被部署在每个块中的4块先后在20天连续5天，。假设只有6只动物的照片捕获，并确定，输入文件1SPACECAP看起来像：
The first data row tells us that Animal ID no 4 was captured at Location ID 1 on the 17th sampling occasion. In your spreadsheet application, you can build up the animal capture details data as per the above format (please make sure that you INCLUDE the header row with the exact titles on column headings as shown above). The file must eventually be saved as a .csv file (for example, "captures.csv") and saved in the working directory. It must be ensured that all Animal ID numbers must be sequenced such that no number is missed in the input file.
第一个数据行告诉我们，动物ID没有被抓住了在17个采样场合地点ID 1。在您的电子表格应用程序，你可以建立动物捕捉细节数据，按上面的格式（请确保您包括标题行中使用确切的标题栏标题如上图所示）。最终文件必须被保存为。csv文件（例如，“captures.csv”），并保存在工作目录。必须确保所有的动物ID号码必须进行测序，没有数字输入文件中的遗漏。
INPUT FILE 2: Trap Deployment Details
INPUT FILE 2：陷阱部署详情
In many camera trap sample surveys of animals, all camera trap stations in the study area may not be operational simultaneously for logistical reasons (for example: limited number of cameras or manpower). Therefore, the trap deployment details input file provides SPACECAP with the information on the dates when each camera trap location was active and operational during the survey. Some trap-nights or trap-days of capture data may be "lost" as a result of camera trap failure, theft, vandalism or animal-damage. This type of trap activity/passivity information can also be effectively fed into and used in SPACECAP. The trap deployment details file records both these types of information, thus accurately accounting for trapping effort.
在许多相机陷阱动物的抽样调查，在研究区所有的相机陷阱站投入运营，同时为后勤方面的原因（例如：相机或数量有限的人力）。因此，陷阱部署的详细信息输入文件提供了SPACECAP的信息时，每个相机陷阱的位置是在调查过程中积极和运营的日期。一些陷阱晚天或陷阱捕获的数据可能会被“丢失”的相机陷阱失败，盗窃，破坏或动物伤害的结果。这种类型的陷阱活动/被动信息也可以被有效地送入和使用在SPACECAP。陷阱部署的详细信息文件记录这两种类型的信息，从而准确地占诱捕效果。
The trap deployment data are stored in a two dimensional matrix of camera trap locations and sampling occasions in a binary, 0/1 format, where 0 indicates that a particular camera trap station was NOT operational on a particular sampling occasion, and 1 indicates that it was operational. The trap location is denoted in 3 columns in the table representing the Trap Location ID no., the spatial location expressed in X and Y-coordinates (in Universal Transverse Mercator UTM projection system in GIS). It is important that these coordinates are represented in the UTM projection system, because it is used for all distance measurements and computations in SPACECAP.
陷阱部署的数据被存储在一个二维矩阵，在一个二进制的摄影机拍摄的位置和采样场合，0/1的格式，其中0表示一个特定的摄影机拍摄的站是一个特定的采样场合NOT上工作，1表示，它是可操作性。陷阱位置表示3列的表中的陷阱位置身份证号码，表示在X和Y坐标（UTM在通用横轴墨卡托投影系统，GEO信息系统）的空间位置。重要的是，这些坐标被表示在UTM投影系统，因为它是用于所有距离的测量和计算，在SPACECAP。
We illustrate trap deployment data entry using a Table for the same example of animal capture details described above. Please recall that the camera trapping survey was conducted with a camera trap array of 16 trap locations, deployed in 4 blocks with 4 camera trap sites in each block. The camera trap survey was conducted over 20 sampling occasions, during which each trapping block was sampled over 5 sampling occasions. The resulting TRAP DEPLOYMENT DATA file would look like the one below:
我们说明陷阱部署数据输入表的动物捕捉细节上述同样的例子。请记得，摄像头捕获的调查进行了部署在4个街区的16个陷阱的位置，在每块有4个摄像头陷阱网站用相机陷阵。相机陷阱进行了调查，在此期间，20个样本的每个捕获块取样超过5个采样场合，。由此产生的TRAP部署数据文件看起来像下面这样：
The table shows that the camera trap sites (Loc 1-4) were operational during sampling occasions 1-5 but were not operational on the remaining sampling occasions 6-20. Additionally, an odd '0' corresponding to Loc 3 and sampling occasion 3 indicates that a trap night was "lost" here.
该表显示，相机陷阱网站（禄1-4）操作期间采样的场合1-5，但不可操作上的其它采样场合6-20。此外，一个奇怪的“0”对应于禄3和采样场合3表明，一个陷阱之夜“丢失”在这里。
The trap deployment details file should be constructed exactly as above in your spreadsheet application. It must then be converted to a comma separated ASCII file (.csv file) with an appropriate filename (e.g. traps .csv) in the working directory accessed by SPACECAP.
陷阱的部署细节文件应当建立，就像上面那样在您的电子表格应用程序。然后，它必须被转换为一个逗号分隔的ASCII文件（。csv文件）用适当的文件名（例如，陷阱。CSV）在工作目录访问SPACECAP。
INPUT FILE 3: Potential Home-Range Centers
输入文件：潜在的家庭型中心
In SPACECAP analyses, the surveyed area containing the camera trap array combined with an extended area surrounding it, known as the "state-space" of the underlying point process, say S, which is represented by a large number of equally spaced points in the form of a very fine mesh. These points are visualized as representing all possible potential activity centers (or home range centers) of all the animals in the animal population being surveyed. This fine grid or mesh of points can be easily generated using a GIS software (like ArcView, MAPINFO etc.) as briefly described below. We view this as an approximation to an underlying continuous state-space which, in practice, would normally be difficult to characterize for computational purposes except in very basic situations where regular polygons might be reasonable. While estimates of population size, N, will be sensitive to the size and extent of the state-space, the estimated density D=N/||S|| is invariant as the extent of the state-space increases. Thus, S should be chosen to be sufficiently large so as to ensure stability of the density estimate. Conceptually, this occurs (under the models fitted in SPACECAP) by choosing S to buffer the trap array by 2 or 3 times the encounter probability scale parameter.
在SPACECAP分析，被调查的区域，其中包含了相机陷阵，围绕着它的扩展区域，被称为“状态空间”的基本观点，说S，表示了大量的等距隔开的形式的很细的网格点。这些点是可视化的，代表所有可能的潜在活动中心（或家庭范围中心）的所有动物在动物种群被调查。这点的精细网格或网可以很容易地使用的GEO信息系统软件（如ArcView的，MAPINFO等）下面简要介绍。我们认为这是一个潜在的连续状态空间，在实践中，通常是难以计算的目的，除了在非常基础的的正多边形的情况下，可能是合理的特点是一个近似。虽然人口规模，N，估计将敏感的状态空间的规模和程度，估计密度D = N / | | S | |状态空间的增加程度是不变的。因此，S的选择应足够大，以确保稳定的密度估计。从概念上讲，这种情况发生时（根据安装在SPACECAP）的模型，通过选择S以缓冲的陷阱的相遇概率尺度参数阵列的2或3倍。
First a rectangle is formed by connecting the outermost camera trap locations using the GIS software of your choice. This rectangle is called the "Minimum area rectangle". A buffer distance (which is sufficiently large to ensure that no individual animal outside of the buffered region has any probability of being photo-captured by the camera traps in the array during the survey) is added to the rectangle around encompassing the trap array. Thereafter, using GIS numerous equally spaced points representing home range centers are generated for this extended area. In practice, some of these potential home range centers in the mesh may end up in habitats known to be entirely unsuitable for the study species (say in the middle of a village, for tiger data). SPACECAP appropriately deals with this problem because "Grid Cells" input file clearly specifies which of these potential home range/activity centers lie within suitable habitat and which do not.
首先，您所选择的GEO信息系统软件的使用，通过连接最外层的相机陷阱的位置形成一个矩形。此矩形被称为“最小面积矩形”。甲缓冲距离（这是足够大，以确保没有个别动物以外的缓冲区域有任何照片拍摄由照相机阵列中的陷阱在统计调查的概率）被添加到周围的矩形包围的陷阱阵列。此后，利用GIS众多代表的活动范围中心的等距点生成的为这个扩展区。在实践中，这些潜在的活动范围中心的网，最终可能已知研究物种（在中间的一个村庄，说为老虎的数据）是完全不适合的生境中。 SPACECAP适当地处理这个问题，因为“网格单元”输入文件明确规定，这些潜在的家庭范围/活动中心内合适的栖息地，没有。
The potential home-range centers data file essentially consists of a 3 column table. The first two columns are the X and Y coordinates (both in UTM projection system) of all the potential activity centers (the equally spaced points generated from your GIS software) and the third column is a habitat suitability indicator column, indicated with 1s if the potential activity centers lies within suitable species habitat or with 0s otherwise. The format of this file is as given below (this example is a small subset of the actual data set, purely for illustration):
潜在的家庭范围中心的数据文件主要由3列的表格。前两列是X和Y坐标（UTM投影系统）的所有潜在的活动中心（您的GIS软件生成的等距点），第三列是一个栖息地适宜度分析指标列，如果用1潜在的活动中心是在合适的物种的栖息地或0，否则。这个文件的格式如下（这个例子是一个小的子集的实际数据集，纯粹是为了说明）：
Example of potential home-range centers input file format:
潜在的家庭范围的中心输入文件的格式示例：
Such a table must be created for all the potential activity centers (this will be a very large table) and saved as a comma delimited ASCII file (.csv file) with an appropriate name, for example, centers.csv, and saved in the working directory.
所有潜在的活动中心，必须创建这样的表（这将是一个非常大的表），并保存为以逗号分隔的ASCII文件（。csv文件）以适当的名称，例如，centers.csv，并保存在工作目录。
In the input data panel, there is a text box for specifying the area of potential home-range centers. This area may be imagined to be composed of a point at the centre of a square, which we call "pixel". Please enter the pixel size area used by the GIS software in this box. The "fineness of the mesh" determined by the pixel size used for spacing of potential home range centers is dictated by species biology (e.g. a few hundred meters for tigers but only a few meters for civets). Caution should be exercised here to not specify a state-space that is too fine because the MCMC algorithm run time increases linearly with the size of the state-space grid. Regarding this as a discrete approximation to some underlying continuous state-space probably justifies a reasonably coarse state-space grid. Regardless of the desired state-space dimension, we recommend carrying out a trial run with a coarse grid to evaluate the performance.
在输入数据“面板中，有一个文本框，用于指定区域的潜在的家庭范围的中心。在一个正方形的中心，我们称之为“像素”组成的一个点，这个区域可能是可想而知的。请使用GIS软件在此框中输入的像素大小的区域。 “成色”确定的网格间距潜在的活动范围中心的像素大小是决定物种生物学（如几百米的老虎，但只有几米果子狸）。应小心不指定状态空间太细，因为的MCMC算法运行时间的增加而线性的状态空间网格的大小。这是一些基本的连续状态空间的离散近似，可能证明一个合理粗糙的状态空间网格。无论所期望的状态空间尺寸，我们建议进行试运行，用粗网格的性能进行评估。
Please load these files using on screen 'buttons' provided in the input data panel of SPACECAP. Please specify the pixel size area of a potential home-range center (as created in input file 3 above) in square kilometers and then click on "OK". Please check the frame at the bottom for status or error messages. In case you wish to edit your selection, please click on the "Edit" button and start the selection all over again.
请加载这些文件上使用的屏幕“按钮，输入数据中的面板SPACECAP。请一个潜在的家庭平方公里范围的中心（如创建在输入上述文件3）在指定像素大小的区域，然后单击“OK”。请检查框架的底部的状态或错误消息。如果你想编辑您的选择，请点击“编辑”按钮，并重新启动的选择。
Proceed to STEP 2, the Model Definition frame.
继续执行第2步，模型定义帧。
STEP 2: SELECTING THE APPROPRIATE MODEL COMBINATION FOR ANALYSIS
第2步：选择合适的模式组合分析
The Model Definition panel of SPACECAP consists of a set of options to select an appropriate model combination for the Spatial-Capture Recapture Analysis. These are simple radio buttons indicating each model choice. Some of these model options that are "grayed out" are expected to be made available in future developments of SPACECAP.
SPACECAP模型定义面板由一组选项中选择一个合适的模式组合的空间捕捉夺回分析。这些都是简单的单选按钮，表示每个模型的选择。这些模式的选择，“灰色”预计将在未来发展SPACECAP。
The model choices are:
该模型的选择是：
1. Trap response present OR Trap response absent Selecting "Trap response present " option runs the behavioral response option (equivalent to Model "Mb"). Select the "Trap response absent" option if you decide otherwise. This model implements a local or "trap-specific" behavioral response under which the probability of encounter in a trap increases subsequent to initial capture in that trap. This is in contrast to the conventional "global" behavioral response which parameterizes a constant increase in encounter probability (on the logit scale, usually) that is not trap specific.
1。出席或陷阱陷阱响应响应缺席，选择：“陷阱响应目前”选项执行的行为反应选项（相当于模型“MB”）。如果你决定，否则，选择“陷阱缺席”选项。这种模式实现了一个本地或“特定陷阱”的行为反应的概率中遇到的一个陷阱，在这个陷阱的初始捕获后增加的。这是在传统的“全球性”的行为反应进行参数化中遇到的概率在不断提高（上的罗吉特规模，通常情况下），不捕获特定的对比。
2. Spatial Capture-Recapture OR Non-spatial Capture-Recapture Select "Spatial Capture-Recapture" for running a spatially explicit capture-recapture analysis, or "Non-spatial Capture-Recapture" for running a conventional capture-recapture analysis (this is equivalent to the Null Model "Mo" in non-spatial CR analysis)
2。空间捕获 - 再捕获或非空间的捕获 - 再选择“空间捕获 - 再捕获运行的空间明确的捕获 - 再捕获分析”，或“非空间捕获 - 再捕获”运行传统的捕获 - 再捕获分析（这是相当于空模型“墨子”非空间CR分析）
3. Half Normal OR Negative Exponential Currently SPACECAP analyses SECR models with only the Half-Normal detection function.
3。半正常或阴性指数目前SPACECAP分析SECR模型只用一半的常规检测功能。
4. Bernoulli (binary) OR Poisson encounter process Currently the analysis is run with the Bernoulli encounter model in which the probability of success is derived as the probability of a positive response under a Poisson encounter rate model. This motivates use of the complementary log-log link which relates encounter probability to distance and other covariates. After the model definition is complete, please click on "OK". Please check the frame at the bottom for status or error messages. In case you wish to edit your selections, please click on the "Edit" button, change your model definition and click on "OK" again. Proceed to the MCMC simulation settings frame.
4。伯努利（二进制）或泊松遇到的过程与目前运行分析中得到成功的概率的概率的积极响应下的泊松遇敌率模型的的伯努利相遇模型。这促使使用的互补性log记录的相遇概率距离和其他变量的链接。在模型定义完成后，请单击“OK”。请检查框架的底部的状态或错误消息。如果你要编辑您的选择，请单击“编辑”按钮，改变你的模型的定义，并再次点击“OK”。进行MCMC模拟设置帧。
STEP 3: Setting the Markov-Chain Monte Carlo (MCMC) parameters (for advanced users)
第3步：设置马尔可夫链蒙特卡罗（MCMC）参数（适用于高级用户）
SPACECAP, uses the Markov-Chain Monte Carlo simulation algorithm written in Program R (R Development Core Team ) to estimate the parameters of the Spatially-explicit Capture Recapture models of Royle et al 2009). The relevant settings can be set in the MCMC simulation settings panel of SPACECAP.
SPACECAP，使用Markov链Monte Carlo模拟算法编写的程序R（R发展的核心小组）估计的参数的空间显式捕获再捕获模型的罗伊尔等人2009）。 MCMC模拟设置面板的SPACECAP中可以设置相关的设置。
No of iterations - This defines the number of MCMC iterations for the analysis (if you aren't sure what this means please set this to a value of about 50,000)
的迭代 - 这定义的数量分析（MCMC迭代，如果你不知道这意味着什么，请设置此值约50,000）
Burn-in - This defines the number of initial values to discard during the MCMC analysis. Setting this at about 1000 usually works well in our experience. Note that some evaluation of whether this is sufficient should be carried out using conventional methods (a topic we will address in subsequent releases).
老化 - 这定义的初始值丢弃在MCMC分析。通常设置在约1000在我们的经验。请注意，应进行一些，这是否是足够的评价利用传统的方法（我们将解决在以后的版本的主题）。
Thinning - This defines the thinning rate. Only iteration numbers defined by the thinning rate are stored during the analysis (if you aren't sure of what this means, you may set this at a value of 1 -that is, no thinning. If you are setting it to 1, please make sure that the number of iterations specified is less than 65,000. This is because Microsoft Excel, which in most versions has a maximum limit of 65000 rows, will not be able to open our output file).
“疏除 - 这定义的稀疏率。只有迭代次数减薄率的定义都存储在分析过程中（如果你不知道这是什么意思，你可以设置此值1，也就是说，没有细化。如果你将它设置为1，请确保指定的迭代数小于65,000，这是因为Microsoft Excel中，在大多数版本的最高限额为65000行，将无法打开输出文件）。
Data augmentation - Since we are uncertain about the total number of animals, which is likely to be larger than the minimum number caught during your camera trapping survey, you will need to "augment" this value by a certain amount. Ideally, you would like this to be a very large number relative to the number you have caught, but setting it up to be very high will cause the analysis to run for a very long time. As a rule of thumb, you could set this to a value of about 5-10 times the number of animals you have photo-captured during the survey. Data augmentation (DA) is a computational device that enables a convenient Bayesian analysis of capture-recapture models where N is unknown. In the context of SECR models, N is the population of individuals having their activity centers on the prescribed state-space. The basic idea of DA is to provide an upper bound on N, say M, which is equal to N plus the number of augmented individuals. Technically, M is the upper limit of a uniform(0,M) prior for N, which is a customary "noninformative" prior for N in this context. As a practical matter, data augmentation creates a list of pseudo-individuals that are always available for the MCMC algorithm to "use" if necessary. That is, these pseudo-individuals leave and enter the population depending on the current values of the model parameters. See Royle, Dorazio and Link (2007) for some general context and Royle et al. (2009) for details in the context of spatial capture-recapture models.
数据增强 - 由于我们不确定总数的动物，这是可能的最低数量大于夹在您的相机捕获的调查，你会需要一定量的“增强”价值。理想情况下，你想这是一个非常大的数目相对于你抓住的数量，但它是非常高的，会导致运行很长一段时间的分析。作为一个经验法则，你可以设置此值约5-10次的动物数量，你有照片在调查过程中捕获。数据增强系统（DA）是一个计算设备，使一个方便的捕获 - 再捕获模型的贝叶斯分析，其中N是未知的。 SECR模型的背景下，N的人口是他们的活动中心在规定的状态空间。对DA的基本思想是提供对N的上限，例如M，这是等于N加增强个人的数目。从技术上说，M是一个统一的上限为N，这是一种习惯“无信息”之前在此上下文中的N（0，M）之前。作为一个实际问题，数据增强的伪MCMC算法“使用”如有必要，随时为您提供的个人创建一个列表。也就是说，这些伪个人离开，根据模型参数的当前值输入人口。请参阅罗伊尔，Dorazio及链接（2007年），为一些大背景下，罗伊尔等的。 （2009）的空间捕获 - 再捕获模型的背景下。
After the MCMC simulation values have been specified, please click on "OK". Please check the frame at the bottom for status or error messages. In case you wish to edit these settings, please click on the "Edit" button, edit these values and click on "OK" again. You are now all set to start the analysis.
MCMC模拟值已被指定后，请点击“OK”。请检查框架的底部的状态或错误消息。如果你想编辑这些设置，请单击“编辑”按钮，编辑这些值，再次点击“OK”。你现在都开始分析。
STEP 4: Running the analysis
STEP 4：运行分析
The last step will simply involve activating the RUN option in the top menu bar. This will start performing the analysis and you will see a progress bar indicating the status of the analysis. Currently, an analysis involving 50000 iterations takes about 14 hours on a fast computer. We are working hard to make the algorithms run faster - we promise!
最后一步将只涉及在顶部的菜单栏激活“运行”选项。这将开始进行分析，你会看到一个进度条指示状态的分析。目前，分析，涉及50000迭代大约需要14个小时，一个快速的电脑上。我们正在努力使算法的运行速度更快 - 我们的承诺！
Results
结果
The posterior density estimates along with standard errors appear as a table in the output panel upon the completion of the analysis. This table also reports estimates of parameters lam0, sigma, psi and beta. If the analysis was run with trap response present, the estimates of "p1" and "p2" are also reported.
后验概率密度估计随着标准表的分析完成后，在输出面板出现错误。此表还报告估计，适马，PSI和βlam0参数。如果分析运行存在的陷阱响应，“P1”和“P2”的估计也有报道。
Additionally, all the results are written into a comma separated file (called <br> param_val_<timestamp>.csv) and is saved into the current working directory. All the summary statistics are written into a file called summary_stats_<timestamp>.csv, which is also saved to the current directory. And the posterior density graphs of all parameters are all stored as jpeg files (.jpg) in the current working directory. SPACECAP also reports a file to generate surface density map (called pixeldensities_val_<timestamp>.csv). This table reports estimates of pixel densities, and the corresponding X_COORD and Y_COORD of the pixels. The table can be then imported into any GIS platform to view the pixel surface densities.
此外，所有结果都写入到一个逗号分隔的文件（称为，参考param_val_的<timestamp>。CSV），并保存到当前的工作目录。所有的汇总统计数据被写入到一个名为summary_stats_的。csv，这也将保存到当前目录中的<timestamp>。后验概率密度图形的所有参数都存储为JPEG文件（JPG格式）在当前工作目录。 SPACECAP还报告了一个文件，以生成表面密度图（pixeldensities_val_的<timestamp>的。csv）。此表报告估计的像素密度，和相应的像素的X_COORD和Y_COORD。该表可以任何GIS平台，查看像素的面密度，然后将其导入。
SPACECAP assesses the convergence of the MCMC run by using the Geweke (1992) diagnostic statistic which is estimated for all the estimated parameters. This statistic produces the z-score values so that a value of |z-score|>1.6 will imply that the MCMC analysis has not been run long enough. SPACECAP also assesses the adequacy of a model using the Bayesian P-value, as implemented in Royle et al. (2011) so that any value that is close to 0 or 1 would imply that the model is inadequate.
SPACECAP通过使用Geweke（1992）可用于所有的估计参数估计的诊断统计评估的收敛的MCMC运行。这一统计数字生产的Z-score值，值| Z-值> 1.6将意味着，MCMC分析尚未运行足够长的时间。 SPACECAP还评估了充足的模型，使用贝叶斯P-值，罗伊尔等实施。 （2011），使任何接近0或1的值，该值是将意味着，该模型是不足的。
Further Developments in progress
取得进展的进一步发展，
We plan to add a number of additional utility modules for various applications, which we will be incorporating in the subsequent versions to be released in next few months. Some of the features we are currently working on are:
我们的各种应用，我们将在随后发布的版本将在未来数月内将计划添加更多的实用工具模块。的功能，我们目前正在为：
1. Include the Poisson encounter process model. <br> 2. Provide with the Hazard-Rate distance function model. <br> 3. Incorporate sex-specific encounters. <br>
1。包括泊松遭遇过程模型。参考2。提供的危险率距离函数模型。 <BR> 3。纳入性别特异性的遭遇。参考
Suggested Citation:
文献引用：
Arjun M. Gopalaswamy, Jeffrey A. Royle, James E. Hines, Pallavi, Singh, Devcharan Jathanna, N. Samba Kumar, K. Ullas Karanth (2011). SPACECAP: A Program to Estimate Animal Abundance and Density using Bayesian Spatially-Explicit Capture-Recapture Models. Wildlife Conservation Society - India Program, Centre for Wildlife Studies, Bengaluru, India. Version 1.0.6.
阿琼M. Gopalaswamy，杰弗里A.罗伊尔，詹姆斯·E.海因斯，Pallavi，辛格，Devcharan Jathanna，K.乌拉斯Karanth（2011年），N.桑巴库马尔。 SPACECAP：A计划，以动物丰度和密度使用空间显式捕获 - 再捕获模型的贝叶斯估计。野生动物保护协会 - 印度，班加罗尔，印度野生动物研究中心计划，。版本1.0.6。
（作者）----------Author(s)----------
Arjun M. Gopalaswamy, Jeffrey A. Royle, James E. Hines, Pallavi Singh, Devcharan Jathanna, N. Samba Kumar and K. Ullas Karanth
参考文献----------References----------
Biometrics 64:377-385.
Statistics, 4(Bernardo, J. M. et al., Eds), 169-193. Oxford University Press.
captures and recaptures. Ecology 79:2852-2862.
vs. web-based density estimators. Ecological Monographs 73:1-26.
Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org.
augmentation. Journal of Computational and Graphical Statistics 16:67-85.
recapture analysis of camera trapping data. Journal of Mammalogy 84:607-614.
Journal of Wildlife Management 49:675-678.
Journal of Mammalogy 66:13-21.
实例----------Examples----------
SPACECAP()
转载请注明:出自 生物统计家园网(http://www.biostatistic.net)。
注：
注1：为了方便大家学习，本文档为生物统计家园网机器人LoveR翻译而成，仅供个人R语言学习参考使用，生物统计家园保留版权。
注2：由于是机器人自动翻译，难免有不准确之处，使用时仔细对照中、英文内容进行反复理解，可以帮助R语言的学习。
注3：如遇到不准确之处，请在本贴的后面进行回帖，我们会逐渐进行修订。
|
发表于 2012-9-30 12:11:35
